libavformat/idcin.c
4120a53a
 /*
ffa5ed24
  * id Quake II CIN File Demuxer
4120a53a
  * Copyright (c) 2003 The ffmpeg Project
  *
b78e7197
  * This file is part of FFmpeg.
  *
  * FFmpeg is free software; you can redistribute it and/or
4120a53a
  * modify it under the terms of the GNU Lesser General Public
  * License as published by the Free Software Foundation; either
b78e7197
  * version 2.1 of the License, or (at your option) any later version.
4120a53a
  *
b78e7197
  * FFmpeg is distributed in the hope that it will be useful,
4120a53a
  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  * Lesser General Public License for more details.
  *
  * You should have received a copy of the GNU Lesser General Public
b78e7197
  * License along with FFmpeg; if not, write to the Free Software
5509bffa
  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
4120a53a
  */
 
 /**
ba87f080
  * @file
ffa5ed24
  * id Quake II CIN file demuxer by Mike Melanson (melanson@pcisys.net)
  * For more information about the id CIN format, visit:
4120a53a
  *   http://www.csse.monash.edu.au/~timf/
  *
  * CIN is a somewhat quirky and ill-defined format. Here are some notes
  * for anyone trying to understand the technical details of this format:
  *
  * The format has no definite file signature. This is problematic for a
  * general-purpose media player that wants to automatically detect file
  * types. However, a CIN file does start with 5 32-bit numbers that
  * specify audio and video parameters. This demuxer gets around the lack
  * of file signature by performing sanity checks on those parameters.
  * Probabalistically, this is a reasonable solution since the number of
  * valid combinations of the 5 parameters is a very small subset of the
  * total 160-bit number space.
  *
  * Refer to the function idcin_probe() for the precise A/V parameters
  * that this demuxer allows.
  *
  * Next, each audio and video frame has a duration of 1/14 sec. If the
  * audio sample rate is a multiple of the common frequency 22050 Hz it will
  * divide evenly by 14. However, if the sample rate is 11025 Hz:
  *   11025 (samples/sec) / 14 (frames/sec) = 787.5 (samples/frame)
  * The way the CIN stores audio in this case is by storing 787 sample
  * frames in the first audio frame and 788 sample frames in the second
  * audio frame. Therefore, the total number of bytes in an audio frame
  * is given as:
  *   audio frame #0: 787 * (bytes/sample) * (# channels) bytes in frame
  *   audio frame #1: 788 * (bytes/sample) * (# channels) bytes in frame
  *   audio frame #2: 787 * (bytes/sample) * (# channels) bytes in frame
  *   audio frame #3: 788 * (bytes/sample) * (# channels) bytes in frame
  *
ffa5ed24
  * Finally, not all id CIN creation tools agree on the resolution of the
4120a53a
  * color palette, apparently. Some creation tools specify red, green, and
  * blue palette components in terms of 6-bit VGA color DAC values which
  * range from 0..63. Other tools specify the RGB components as full 8-bit
  * values that range from 0..255. Since there are no markers in the file to
  * differentiate between the two variants, this demuxer uses the following
  * heuristic:
  *   - load the 768 palette bytes from disk
  *   - assume that they will need to be shifted left by 2 bits to
  *     transform them from 6-bit values to 8-bit values
  *   - scan through all 768 palette bytes
  *     - if any bytes exceed 63, do not shift the bytes at all before
  *       transmitting them to the video decoder
  */
 
6a5d31ac
 #include "libavutil/intreadwrite.h"
4120a53a
 #include "avformat.h"
 
 #define HUFFMAN_TABLE_SIZE (64 * 1024)
40e8e497
 #define IDCIN_FPS 14
4120a53a
 
 typedef struct IdcinDemuxContext {
     int video_stream_index;
     int audio_stream_index;
     int audio_chunk_size1;
     int audio_chunk_size2;
 
     /* demux state variables */
     int current_audio_chunk;
     int next_chunk_is_video;
     int audio_present;
 
     int64_t pts;
 } IdcinDemuxContext;
 
 static int idcin_probe(AVProbeData *p)
 {
     unsigned int number;
 
     /*
ffa5ed24
      * This is what you could call a "probabilistic" file check: id CIN
4120a53a
      * files don't have a definite file signature. In lieu of such a marker,
      * perform sanity checks on the 5 32-bit header fields:
      *  width, height: greater than 0, less than or equal to 1024
      * audio sample rate: greater than or equal to 8000, less than or
      *  equal to 48000, or 0 for no audio
      * audio sample width (bytes/sample): 0 for no audio, or 1 or 2
      * audio channels: 0 for no audio, or 1 or 2
      */
 
8466ab59
     /* check we have enough data to do all checks, otherwise the
        0-padding may cause a wrong recognition */
     if (p->buf_size < 20)
         return 0;
 
4120a53a
     /* check the video width */
fead30d4
     number = AV_RL32(&p->buf[0]);
4120a53a
     if ((number == 0) || (number > 1024))
        return 0;
 
     /* check the video height */
fead30d4
     number = AV_RL32(&p->buf[4]);
4120a53a
     if ((number == 0) || (number > 1024))
        return 0;
 
     /* check the audio sample rate */
fead30d4
     number = AV_RL32(&p->buf[8]);
4120a53a
     if ((number != 0) && ((number < 8000) | (number > 48000)))
         return 0;
 
     /* check the audio bytes/sample */
fead30d4
     number = AV_RL32(&p->buf[12]);
4120a53a
     if (number > 2)
         return 0;
 
     /* check the audio channels */
fead30d4
     number = AV_RL32(&p->buf[16]);
4120a53a
     if (number > 2)
         return 0;
 
     /* return half certainly since this check is a bit sketchy */
     return AVPROBE_SCORE_MAX / 2;
 }
 
 static int idcin_read_header(AVFormatContext *s,
                              AVFormatParameters *ap)
 {
471fe57e
     AVIOContext *pb = s->pb;
e4141433
     IdcinDemuxContext *idcin = s->priv_data;
4120a53a
     AVStream *st;
     unsigned int width, height;
     unsigned int sample_rate, bytes_per_sample, channels;
 
     /* get the 5 header parameters */
e63a3628
     width = avio_rl32(pb);
     height = avio_rl32(pb);
     sample_rate = avio_rl32(pb);
     bytes_per_sample = avio_rl32(pb);
     channels = avio_rl32(pb);
4120a53a
 
     st = av_new_stream(s, 0);
     if (!st)
769e10f0
         return AVERROR(ENOMEM);
40e8e497
     av_set_pts_info(st, 33, 1, IDCIN_FPS);
4120a53a
     idcin->video_stream_index = st->index;
72415b2a
     st->codec->codec_type = AVMEDIA_TYPE_VIDEO;
01f4895c
     st->codec->codec_id = CODEC_ID_IDCIN;
     st->codec->codec_tag = 0;  /* no fourcc */
     st->codec->width = width;
     st->codec->height = height;
4120a53a
 
     /* load up the Huffman tables into extradata */
01f4895c
     st->codec->extradata_size = HUFFMAN_TABLE_SIZE;
     st->codec->extradata = av_malloc(HUFFMAN_TABLE_SIZE);
e63a3628
     if (avio_read(pb, st->codec->extradata, HUFFMAN_TABLE_SIZE) !=
4120a53a
         HUFFMAN_TABLE_SIZE)
6f3e0b21
         return AVERROR(EIO);
4120a53a
 
     /* if sample rate is 0, assume no audio */
     if (sample_rate) {
         idcin->audio_present = 1;
         st = av_new_stream(s, 0);
         if (!st)
769e10f0
             return AVERROR(ENOMEM);
40e8e497
         av_set_pts_info(st, 33, 1, IDCIN_FPS);
4120a53a
         idcin->audio_stream_index = st->index;
72415b2a
         st->codec->codec_type = AVMEDIA_TYPE_AUDIO;
01f4895c
         st->codec->codec_tag = 1;
         st->codec->channels = channels;
         st->codec->sample_rate = sample_rate;
dd1c8f3e
         st->codec->bits_per_coded_sample = bytes_per_sample * 8;
01f4895c
         st->codec->bit_rate = sample_rate * bytes_per_sample * 8 * channels;
         st->codec->block_align = bytes_per_sample * channels;
4120a53a
         if (bytes_per_sample == 1)
01f4895c
             st->codec->codec_id = CODEC_ID_PCM_U8;
4120a53a
         else
01f4895c
             st->codec->codec_id = CODEC_ID_PCM_S16LE;
4120a53a
 
         if (sample_rate % 14 != 0) {
             idcin->audio_chunk_size1 = (sample_rate / 14) *
             bytes_per_sample * channels;
             idcin->audio_chunk_size2 = (sample_rate / 14 + 1) *
                 bytes_per_sample * channels;
         } else {
             idcin->audio_chunk_size1 = idcin->audio_chunk_size2 =
                 (sample_rate / 14) * bytes_per_sample * channels;
         }
         idcin->current_audio_chunk = 0;
     } else
         idcin->audio_present = 1;
 
     idcin->next_chunk_is_video = 1;
     idcin->pts = 0;
 
     return 0;
 }
 
 static int idcin_read_packet(AVFormatContext *s,
                              AVPacket *pkt)
 {
     int ret;
     unsigned int command;
     unsigned int chunk_size;
e4141433
     IdcinDemuxContext *idcin = s->priv_data;
471fe57e
     AVIOContext *pb = s->pb;
4120a53a
     int i;
     int palette_scale;
2a2bbcb0
     unsigned char r, g, b;
     unsigned char palette_buffer[768];
2d8591c2
     uint32_t palette[256];
4120a53a
 
899681cd
     if (url_feof(s->pb))
6f3e0b21
         return AVERROR(EIO);
4120a53a
 
     if (idcin->next_chunk_is_video) {
e63a3628
         command = avio_rl32(pb);
4120a53a
         if (command == 2) {
6f3e0b21
             return AVERROR(EIO);
4120a53a
         } else if (command == 1) {
             /* trigger a palette change */
e63a3628
             if (avio_read(pb, palette_buffer, 768) != 768)
6f3e0b21
                 return AVERROR(EIO);
4120a53a
             /* scale the palette as necessary */
             palette_scale = 2;
             for (i = 0; i < 768; i++)
2a2bbcb0
                 if (palette_buffer[i] > 63) {
4120a53a
                     palette_scale = 0;
                     break;
                 }
 
2a2bbcb0
             for (i = 0; i < 256; i++) {
                 r = palette_buffer[i * 3    ] << palette_scale;
                 g = palette_buffer[i * 3 + 1] << palette_scale;
                 b = palette_buffer[i * 3 + 2] << palette_scale;
2d8591c2
                 palette[i] = (r << 16) | (g << 8) | (b);
2a2bbcb0
             }
4120a53a
         }
 
e63a3628
         chunk_size = avio_rl32(pb);
4120a53a
         /* skip the number of decoded bytes (always equal to width * height) */
45a8a02a
         avio_skip(pb, 4);
4120a53a
         chunk_size -= 4;
115329f1
         ret= av_get_packet(pb, pkt, chunk_size);
044a950d
         if (ret < 0)
             return ret;
2d8591c2
         if (command == 1) {
             uint8_t *pal;
 
             pal = av_packet_new_side_data(pkt, AV_PKT_DATA_PALETTE,
                                           AVPALETTE_SIZE);
             if (ret < 0)
                 return ret;
             memcpy(pal, palette, AVPALETTE_SIZE);
         }
4120a53a
         pkt->stream_index = idcin->video_stream_index;
         pkt->pts = idcin->pts;
     } else {
         /* send out the audio chunk */
         if (idcin->current_audio_chunk)
             chunk_size = idcin->audio_chunk_size2;
         else
             chunk_size = idcin->audio_chunk_size1;
2692067a
         ret= av_get_packet(pb, pkt, chunk_size);
044a950d
         if (ret < 0)
             return ret;
4120a53a
         pkt->stream_index = idcin->audio_stream_index;
         pkt->pts = idcin->pts;
 
         idcin->current_audio_chunk ^= 1;
40e8e497
         idcin->pts++;
4120a53a
     }
 
     if (idcin->audio_present)
         idcin->next_chunk_is_video ^= 1;
 
     return ret;
 }
 
66355be3
 AVInputFormat ff_idcin_demuxer = {
4120a53a
     "idcin",
517ac243
     NULL_IF_CONFIG_SMALL("id Cinematic format"),
4120a53a
     sizeof(IdcinDemuxContext),
     idcin_probe,
     idcin_read_header,
     idcin_read_packet,
 };